Comparison of Reconstruction Algorithm on Sparse Representation based Classification (SRC) for Face Recognition

Abstract

Sparse representation based Classification (SRC) has gained the attention of pattern recognition and computer vision researchers, especially researchers working on face recognition. On SRC's algorithm, it is necessary to find a solution to an optimization problem to recover $\mathbf{x}$ from the equation $\mathbf{y}$ = Ax. Only a few studies reported the reconstruction of the signals on SRC's algorithm. Therefore, this paper studies the comparison of OMP, LASSO, and CVX to help the readers understand the reconstruction algorithm's effect on SRC. The simulation result is that LASSO and CVX algorithms have the same recognition rate, but LASSO can compute twice faster as CVX. On the other hand, the OMP algorithm can give the highest recognition rate on a specific dimension of the image with a faster computation time than LASSO.